Oncogene (2006) 25, 5741–5751 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE CD44 ligation induces -independent cell death via a novel / AIF pathwayin human erythroleukemiacells

C Artus1, E Maquarre1, RS Moubarak2, C Delettre2, C Jasmin1, SA Susin2 and J Robert-Le´ ze´ ne` s1

1INSERM U602, Hoˆpital Paul Brousse, Villejuif, France and 2Apoptose et Syste`me Immunitaire Group, Institut Pasteur, Paris, France

Ligation of the cell surface molecule CD44 byanti-CD44 CD44 with specific anti-CD44 monoclonal antibodies monoclonal antibodies (mAbs) has been shown to induce (mAbs) can reverse differentiation blockage of the cell differentiation, cell growth inhibition and in some leukemia cells in primary blasts from patients with cases, apoptosis in myeloid leukemic cells. We report, AML-M1 to AML-M5 subtypes (Charrad et al., 1999). herein, that exposure of human erythroleukemic HEL More recently, the anti-CD44 mAb HI44a has been cells to the anti-CD44 mAb A3D8 resulted in cell growth reported to induce both differentiation and apoptosis in inhibition followed bycaspase-independent apoptosis-like AML cells (Song et al., 2004). In addition, the two cell death. This process was associated with the disruption activating anti-CD44 mAbs H90 and A3D8 trigger of mitochondrial membrane potential (DWm), the mito- terminal differentiation in several human myeloid chondrial release of apoptosis-inducing factor (AIF), but leukemia cell lines, strongly inhibit their proliferation not of cytochrome c, and the nuclear translocation of AIF. and, in some cases, induce apoptotic cell death (Charrad All these effects including cell death, loss of mitochondrial et al., 2002). Downregulation of c-Jun (Zada et al., 2003) DWm and AIF release were blocked bypretreatment with and c-Myc (Song et al., 2004) expression and upregula- the poly(ADP-ribose) polymerase inhibitor isoquinoline. tion of the cyclin-dependent kinase inhibitor p27kipÀ1 A significant protection against cell death was also (Gadhoum et al., 2004b) have recently been suggested to observed byusing small interfering RNA for AIF. explain the inhibition of cell proliferation upon CD44 Moreover, we show that calpain was activated ligation in AML cells. Recent data show that the anti- before the appearance of apoptosis, and that calpain CD44 mAb A3D8 induces apoptosis via caspase- and inhibitors or transfection of calpain-siRNA decrease serine-protease-dependent pathways in acute promyelo- A3D8-induced cell death, and block AIF release. These cytic leukemia NB4 cells (Maquarre et al., 2005). data suggest that CD44 ligation triggers a novel caspase- However, the signaling pathway by which anti-CD44 independent cell death pathwayvia calpain-dependent AIF mAbs induce cell death in other AML cells remains release in erythroleukemic HEL cells. undefined. Understanding the mechanism of CD44- Oncogene (2006) 25, 5741–5751. doi:10.1038/sj.onc.1209581; induced apoptosis would help in identifying novel published online 24 April 2006 therapeutic targets potentially useful in clinical trials. Cell death patterns have been divided into pro- Keywords: CD44; apoptosis; calpain; AIF; erythro- grammed cell death (PCD) or apoptosis and accidental leukemia or passive necrotic cell death (Leist and Jaattela, 2001). Apoptosis is actually the best characterized type of PCD: cells display membrane blebbing, loss of the asymmetry of phosphatidylserine (PS) in the plasma membrane, nuclear fragmentation, and activated Introduction , a family of cell-suicide cysteine (Nicholson, 1999). The biochemical activation of CD44 is a cell surface antigen that demonstrates cell apoptosis occurs via two major pathways: the intrinsic adhesion and signaling function (Ponta et al., 2003). pathway, initiated by the mitochondrial release of CD44 is involved in normal myelopoiesis and lympho- cytochrome c, resulting in the activation of caspase-9; poiesis and is also expressed on leukemia blasts in all and the extrinsic pathway, initiated by the cell surface acute myeloid leukemia (AML) subtypes (Ghaffari death receptors such as Fas that leads to the activation et al., 1999). It has been reported that the ligation of of caspase-8 or -10 (Sun et al., 1999). Both pathways converge and activate executioner caspase-3, which then cleaves intracellular substrates and causes Correspondence: Dr J Robert-Le´ ze´ ne` s, INSERM U602, Hoˆ pital Paul- cell death. There is now accumulating evidence indi- Brousse, 14 avenue Paul Vaillant Couturier, 94807 Villejuif cedex, cating that PCD can also occur in complete absence France. E-mail: [email protected] and independently of caspase activation (Johnson, Received 27 October 2005; revised 16 February 2006; accepted 8 March 2000; Jaattela and Tschopp, 2003; Lockshin and 2006; published online 24 April 2006 Zakeri, 2004; Broker et al., 2005). Several models of CD44-induced cell death in erythroleukemia cells CArtuset al 5742 caspase-independent PCD have been described. They Results include autophagy, paraptosis, mitotic catastrophe or the descriptive models of apoptosis-like and necrosis- Effect of CD44 ligation with A3D8 antibody in HEL cells like PCD (Broker et al., 2005). Caspase-independent By using A3D8 anti-CD44 mAb, preliminary results cell death pathways are important mechanisms for showed that HEL cells expressed CD44 as evaluated by triggering a response to cytotoxic agents or other death flow cytometry analysis (data not shown). When A3D8 stimuli when the caspase-mediated routes fail. These mAb was added to HEL cells for 4 days, a pronounced pathways may involve organelles such as lysosomes inhibition of cell proliferation was observed (Figure 1a, and the endoplasmic reticulum via the activation of upper panel). When two other anti-CD44 H90 and J173 non-caspases proteases such as or mAbs were tested, no such effects were observed (data (Johnson, 2000). However, the mitochondrion plays a not shown). The inhibitory effect of A3D8 was time- central role in caspase-independent cell death as in and-concentration-dependent. Thus, the decrease of cell caspase-dependent apoptosis (Bras et al., 2005). number appeared within 1–2 days and was most Key events in classic apoptosis as well as in other pronounced at high concentrations of A3D8 (3 mg/ml). forms of cell death are the disruption of mitochondrial At this latter concentration, a decrease in viable cells function and the release of apoptogenic from occurred at day 3 concomitantly with appearance of cell the intermembrane space (IMS) of mitochondria death (Figure 1a, lower panel). Cell death was found to (Henry-Mowatt et al., 2004; Saelens et al., 2004). be also dose-dependent and reached 25–30% at 3 mg/ml Although the process of the mitochondrial outer of A3D8 by day 4. With further exposure, the membrane permeabilization (MOMP) is mainly con- proportion of dead cells increased up to about 60% by trolled by the members of the Bcl-2 family, the exact day 5 (data not shown). mechanisms responsible of this process remain contro- We next investigated whether A3D8-induced inhibi- versial (Donovan and Cotter, 2004). Whatever the tion of proliferation was owing to a cell cycle perturba- mechanism required for MOMP, release of mitochon- tion. A marked accumulation of HEL cells in S phase drial proteins causes most forms of caspase-dependent or -independent cell death. Mitochondrial proteins that induce caspase-dependent apoptosis include cytochrome c and two other proteins, Smac/Diablo and Omi/HtrA2, which antagonize the inhibitors of apoptosis (IAPs). Mitochondria can also release EndoG and apoptosis- inducing factor (AIF), one of the major mediators of cell death involved in caspase-independent apoptosis (Joza et al., 2001; Cande et al., 2004; Cregan et al., 2004). Apoptosis-inducing factor is expressed as a precursor of 67 kDa, which is addressed and compartmentalized into mitochondria by two-mitochondrial localization sequences located within the N-terminal prodomain of the protein (Susin et al., 1999). Once in mitochondria, the full-length AIF is processed and the prodomain removed, giving rise to a mature form of B57 kDa (Otera et al., 2005). Under physiological conditions, AIF is a mitochondrial FAD-dependent oxidoreductase that plays a role in oxidative phosphorylation (Miramar et al., 2001). However, after a cellular insult, AIF is cleaved by calpains and/or cathepsins (Polster et al., 2005; Yuste et al., 2005) and translocates from mitochondria to cytosol and to nucleus where it causes, in a caspase-independent fashion, chromatin condensa- tion and large-scale (B50 kb) DNA fragmentation (Susin et al., 1999). In the present study, we investigated the nature of the apoptotic signals mediated by CD44 ligation with the specific A3D8 mAb in human erythroleukemic HEL cells, a model of M6 subtype of AML that is usually refractory to conventional drugs. Our data reveal that Figure 1 Anti-CD44 mAb A3D8 induces cell growth inhibition CD44 ligation by A8D8 induced caspase-independent followed by cell death in HEL cells. (a) HEL cells were seeded at apoptosis-like cell death in HEL cells via AIF transloca- 2 Â 105 cells/ml and treated with various concentrations of A3D8 tion to nuclei and poly (ADP-ribose) polymerase (0.5–3 mg/ml). Control HEL cells were treated with IgG1. Trypan blue test was used to estimate cell growth (top) and cell death (PARP) activation. We also show that calpain activa- (bottom) along 4 days of treatment. Data represent the means7s.d. tion was involved during the apoptotic process and was of triplicate cultures. (b) Cell cycling was analysed in HEL cells required for AIF release and its nuclear translocation. treated with 1 mg/ml of A3D8 or IgG1 (control) for 3 days.

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5743 was observed when treated with 1 mg/ml of A3D8 for 3 A3D8 induces DCmdisruption without cytochrome c days (84% in A3D8 treated cells compared with 40% in release control cells) (Figure 1b). It is noteworthy that such a To determine the putative involvement of mitochon- dose of A3D8 did not induce cell death at day 3 while drial-dependent pathways in A3D8-induced cell death, inhibiting cell growth (see Figure 1a). Therefore, CD44 we examined the DCm by using flow cytometric analysis ligation with A3D8 mAb induced HEL cell accumula- of DiOC6 staining. HEL cell treatment with A3D8 tion in S phase of the cell cycle as well as inhibition of induced a decrease of DiOC6 staining within 2 days, cell proliferation before cell death appeared. and 40% treated cells displayed low DiOC6 staining after 4 days (Figure 4a). Therefore, DCm disruption was observed 24 h before cell death occurred. Because A3D8 induces an apoptosis-like cell death in HEL cells mitochondrial alterations were detected in A3D8- To further characterize the nature of the A3D8-induced treated HEL cells, the release of the two proapoptotic cell death, we used acridine-orange(AO)/ethidium bro- factors cytochrome c and AIF from mitochondria to mide assay(EB) that analysed cell and nucleus morpho- cytosol was examined. Cellular fragmentation experi- logy and allowed to distinguish apoptotic from necrotic ments did not show cytochrome c in the cytosolic cells. After 3- and 4-day treatment with 3 mg/ml of fraction after a 4-day treatment with A3D8 whereas A3D8, HEL cell exhibited features typical of apoptosis, present in etoposide-treated HEL cells (Figure 4b). In including chromatin condensation (Figure 2a). More- contrast to cytochrome c, we observed the release of over, 20.5% of HEL cells were found to be positive for AIF from mitochondria to cytosol in A3D8-treated V (Figure 2b). Apoptosis was further con- HEL cells (Figure 4b). firmed by TUNEL assay that specifically labelled DNA strands brakes and detected DNA fragmentation in HEL cells by day 3 that was increased by day 4 Apoptosis-inducing factor translocates to the nucleus and (Figure 2c). Finally, apoptotic features such as con- mediates A3D8-induced cell death in HEL cells densed chromatin were observed by using electron As A3D8 induced chromatin condensation without microscopy (Figure 2d). Chromatin, however, did not caspase activation but with cytosolic release of AIF, condense to crescent-shaped figures typical of apoptosis we determined whether AIF translocated to the nucleus but rather to lumpy shapes. Moreover, necrotic-like during A3D8 treatment. As shown in Figure 5a, AIF ultrastructural features, including cytoplasmic vesicula- was detected in the nuclear fraction of A3D8-treated tion and mitochondrial swelling with cristae distension, HEL cells at day 4. Nuclear AIF localization in A3D8- were also present in HEL cells treated with A3D8. treated cells was further confirmed by confocal micro- scopy where both cytosolic and nuclear AIF pattern was observed at day 4 whereas AIF was only detected in the Cell death induced by A3D8 is caspase independent cytosol in control HEL cells (Figure 5b). To determine To assess the role of caspases in apoptotic type–cell whether nuclear AIF displayed nuclease activity, DNA death induced by A3D8 mAb, HEL cells were incubated pulsed field gel migration experiments were realized and with A3D8 in presence of the general caspase inhibitor showed an accumulation of 50 kb fragmented DNA at Z-VAD-fmk. As shown in Figure 3a, Z-VAD-fmk at a day 4 in A3D8-treated cells (Figure 5c). This large-scale concentration of 25 mM, that is sufficient to inhibit DNA fragmentation was typical of an AIF activity caspases, did not significantly decrease A3D8-induced (Susin et al., 1999). Finally, AIF involvement in A3D8- cell death. However, Z-VAD-fmk at the higher dose of induced cell death was investigated by using AIF-small 50 mM was able to rescue from death 40% to 50% of interfering RNA (siRNA) transfection. As illustrated in A3D8-treated cells for 4 days (Figure 3a). This can be Figure 5d, AIF-siRNA was efficient to decrease AIF explained by the fact that, when used at these higher protein levels in HEL cells as compared with cells doses, Z-VAD-fmk is no longer specific and could also transfected with an irrelevant control-siRNA (Ctrl- inhibit other cysteine proteases, like calpains or cathe- siRNA). A significant reduction of cell death was psins (Schotte et al., 1999). In contrast, 50 mM of Z- observed in A3D8-treated cells transfected with AIF- DEVD-fmk, a specific inhibitor of the effector caspase-3 siRNA whereas absent in cells transfected with Ctrl- and -7, was completely inefficient in inhibiting cell death siRNA (Figure 5d). These results strongly suggest that induced by A3D8 after 4 days of treatment (Figure 3a). AIF is involved in HEL cell death induced by CD44 Similarly, specific inhibitors of caspase-6 (Z-VEID- ligation with A3D8 mAb. fmk), caspase-2 (Z-VDVAD-fmk) and caspase-1 (Z- YVAD-fmk) did not block A3D8-induced cell death. Western blotting analysis shows that a 4-day treatment Poly (ADP-ribose) polymerase-1 is involved in cell death with A3D8 failed to induce cleavage of the executioner induced by A3D8 and in apoptosis-inducing factor nuclear caspase-3 and caspase-7, and initiator caspase-8 and translocation caspase-9, whereas it is present in HEL cells treated with Recent studies indicate that AIF is a key mediator of etoposide (Figure 3b). Likewise, A3D8 treatment was caspase-independent cell death mediated by PARP-1 unable to induce the cleavage of the major caspase-3/-7 activation (Yu et al., 2002; Hong et al., 2004). We target PARP-1 (Figure 3c). Taken together these data therefore tested the effects of pharmacological PARP-1 suggest that CD44 ligation with A3D8 induced a inhibitor 1,5-isoquinolinediol on A3D8-treated HEL caspase-independent cell death in HEL cells. cells. Co-treatment of cells with A3D8 and PARP-1

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5744

Figure 2 A3D8 induces an apoptosis-like cell death in HEL cells. HEL cells were treated with 3 mg/ml of A3D8 or with IgG1 (control). (a) Detection of apoptotic cells by acridine-orange/ethidium bromide staining. Left panel: fluorescence microscopy examination of stained cells; right panel: percentage (%) of apoptotic cells estimated at indicated days. Values are means7s.d. of triplicate cultures. (b) Flow cytometry analysis of annexin V/PI staining in cells treated for 4 days with IgG1 (control) or A3D8. (c) Percentage (%) of TUNEL-positive cells evaluated by cytometry at indicated days. Values are means7s.d. of triplicate cultures. (d) Electronic microscopy analysis of HEL cells cultured for 4 days with IgG1 (control) or A3D8.

inhibitor severely reduced A3D8-induced cell death way we found that, after A3D8 treatment, NAD þ (Figure 6a) in association with a partially rescue of decreased in a time-dependent manner with kinetics DCm disruption (Figure 6b). As PARP-1 activation similar to A3D8-mediated death response. Indeed, consumes b-nicotinamide-adenine dinucleotide (NAD þ ) NAD þ decreased at day 2 after A3D8 treatment and resulting in a reduction of NAD þ levels, we investigated was reduced to B50% of control levels at day 4. NAD þ levels in A3D8-treated cells (Figure 6c). In this Furthermore, as shown in Figure 6d, PARP-1 inhibitor

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5745

Figure 4 A3D8 induces a stroke of Cm with mitochondrial release of apoptosis-inducing factor (AIF) but not of cytochrome c. (a) HEL cells were treated with IgG1 (control) or 3 mg/ml of A3D8. Cm was detected by flow cytometry using DiOC6 incorporation and the percentage (%) of cells with low Cm is estimated at indicated days. Values are means7s.d. of three independent experiments. *Po0.05 compared with A3D8 alone (b) Cytosolic fractions of cells treated for 4 days ( þ ) or not (À) with A3D8 were analysed by Western blot with anti-cytochrome c and anti-AIF antibodies. HEL cells treated for 36 h with 100 mM of etoposide were used as positive control for cytochrome c release. Membrane was striped and subsequently probed with anti-actin antibody as loading control for cytosolic fraction.

prompted us to investigate the potential role of another . Interestingly, when two calpain inhibitors calpeptin or ALLM were added to HEL cells Figure 3 A3D8 induces a caspase-independent cell death. (a) HEL in the presence of A3D8, 50% inhibition of cell death cells were incubated with IgG1 (control), or A3D8 (3 mg/ml) in the was observed (Figure 7a). This fact suggests the absence or presence of the following inhibitors: pancaspase involvement of calpains in HEL cell death induced by inhibitor Z-VAD-fmk (25 or 50 mM); inhibitor of caspase-3/7 Z- DEVD-fmk (50 mM); inhibitor of caspase-6 Z-VEID-fmk (50 mM); CD44 ligation with A3D8 mAb. This was confirmed by inhibitor of caspase-2 Z-VDVAD (10 mM); inhibitor of caspase-1 three independent methods. First, by using specific YVAD-fmk (20 mM). The percentage (%) of cell death was calpain-siRNAs. As shown in Figure 7b two siRNAs for estimated by trypan blue dye exclusion. Values are means7s.d. calpain small unit (CAPNS1-siRNA1 and CAPNS1- of triplicate cultures. *Po0.05 compared with A3D8 alone. (b) and (c) Total proteins from HEL cells cultured with 3 mg/ml of A3D8 or siRNA2) were efficient to decrease calpain small subunit IgG1 (control) were analysed by SDS-polyacrylamide gel electro- protein levels in HEL cells as compared with cells phoresis at indicated times and immunoblotted with anti-caspase-8, transfected with a negative control-siRNA (NC-siRNA). anti-caspase-9, anti-caspase-3, anti-caspase-7 antibodies (b) and A significant reduction of cell death was observed in anti-poly (ADP-ribose) polymerase-1 antibody (c). HEL cells A3D8-treated cells transfected with both CAPNS1- incubated for 36 h with 100 mM of etoposide were used as positive control. siRNAs whereas absent in cells transfected with NC- siRNA (Figure 7b). Second, by a western immunoblot analysis. This method detected the cleaved forms of the reduced by less than 60% the quantity of AIF detected small subunit of calpain, which are typical of calpain in the nucleus of A3D8-treated cells indicating that activation, in A3D8-treated HEL cells by day 2 PARP-1 activation triggers AIF nuclear translocation. (Figure 7c). Third, by treatment of HEL cells by a Ca2 þ chelator. Indeed, as calcium is required for calpain Calpain is activated by the treatment with A3D8 and activation, pre-incubation with the intracellular Ca2 þ involved in apoptosis-inducing factor nuclear translocation chelator, BAPTA/AM significantly prevented A3D8- As A3D8-induced cell death did not appear to involve induced cell death (Figure 7a). Taken together, these caspases, the effective inhibitory effects of high con- results clearly indicate that calpain is activated and centrations of the caspase inhibitor Z-VAD-fmk involved in CD44-mediated HEL cell death.

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5746

Figure 5 A3D8 promotes an apoptosis-inducing factor (AIF)-dependent cell death in HEL cells. (a) Nuclear fraction of HEL cells treated with 3 mg/ml of A3D8 or IgG1 (control) for 4 days was analysed by Western blot with anti-AIF antibody. Membrane was striped and probed with anti-poly (ADP-ribose) polymerase-1 as loading control for nuclear fraction. (b) Differential subcellular localization of AIF in HEL cells treated for 4 days with A3D8 or IgG1 (control) visualized by confocal microscopy. (c) Large-scale DNA fragment analysed by FIGE in HEL cells treated ( þ ) or not (À) with A3D8. (d) HEL cells were transfected with control-small interfering RNA(siRNA) (Ctrl-siRNA) or AIF-siRNA and 24 h after electroporation were cultured with IgG1 or with A3D8. Cell death was estimated by trypan blue exclusion test at 3 days of A3D8 treatment and compared with not transfected (control) cells. Values are means7s.d. of triplicate cultures. *Po0.05 compared with A3D8 alone or with A3D8 þ Ctrl-siRNA. The inserted Western blot picture shows that nuclear AIF protein was diminished 3 days after electroporation in A3D8-treated cells transfected with AIF- siRNA as compared with Ctrl-siRNA.

Figure 6 Poly (ADP-ribose) polymerase (PARA-1) activation mediates A3D8-induced cell death and apoptosis-inducing factor (AIF) nuclear translocation. HEL cells were incubated with IgG1 (control) or A3D8 (3 mg/ml) in the absence or presence of 10 mM PARP inhibitor 1,5 isoquinolinediol (ISO). Percentage (%) of cell death was estimated by trypan blue dye exclusion after 4 days of incubation (a) and percentage (%) of cells with low Cm was estimated at indicated days (b). Values are means7s.d. of three independent experiments. (c) HEL cells were incubated with IgG1 (control) or A3D8 (3 mg/ml) and NAD þ levels were determined at indicated times. Concentrations of NAD þ were normalized with that of control cells. Values are means7s.d. of three independent experiments. (d) Top: Nuclear fraction of HEL cells cultured for 4 days with IgG1 (control) or A3D8 (3 mg/ml) in the absence or presence of 10 mM ISO was analysed by Western blotting with anti-AIF antibody. Membrane was striped and probed with anti-PARP-1 antibody as loading control for nuclear material. Bottom: Densitometry quantification of changes in nuclear AIF was derived from Western blots of three independent experiments with arbitrary unit corresponding to A3D8-treated cells.

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5747 Finally, and because it has recently reported that followed by cell death in a cellular model of AML-M6 calpain I can induce cleavage and release of AIF from subtype, the erythroleukemic HEL cell line. CD44 isolated mitochondria (Polster et al., 2005), we examined ligation triggers in HEL cells an original form of whether calpain inhibition could affect AIF release apoptosis-like caspase-independent cell death involving observed during A3D8-induced cell death. Treatment both AIF and calpain. Moreover, we show for the first with the calpain inhibitor calpeptin (Figure 7d) effi- time that calpain activation can mediate nuclear AIF ciently decreased AIF nuclear translocation in A3D8- translocation in HEL cells. treated HEL cells. Moreover, as shown in the Figure 7e, It has been reported that several anti-CD44 mAbs the release of AIF from mitochondria to cytosol was induce cell growth inhibition in AML M1-M5 subtypes blocked by CAPNS1-siRNA2 but not by NC-siRNA tested (Charrad et al., 1999). Only A3D8 mAb and not suggesting that calpain activation is involved in the AIF other anti-CD44 J173 and H90 mAbs mediate the release governing this kind of cell death. inhibition of cell proliferation in HEL cells. These differences of activity between the different antibodies may be explained by the fact that they map distinct Discussion epitopes of the CD44 molecule. It is known that the activity of CD44 is dictated by the cell type and the anti- The present report shows that ligation of CD44 with CD44 mAb agonist molecule (Gadhoum et al., 2004a). A3D8 mAb induces an inhibition of proliferation It has been shown that A3D8 mAb induces G1 phase arrest and p27/Kip up-regulation in two myeloid leukemic NB4 and HL-60 cells (Gadhoum et al., 2004b). Contrastingly, we observed that A3D8 mAb induced HEL cell growth arrest in S phase suggesting that the response to CD44 ligation to a same agonist mAb depends on the AML cell subtype. Elucidation of the pathways involved in CD44-mediated S phase arrest in HEL cells will be of future interest. Moreover, the functional contribution, if any, of such cell cycle perturbations to CD44-induced cell death remains to be established. Cell death induced by A3D8 mAb in HEL cells displayed several features typical of apoptosis, including cytological changes (cell shrinkage and nuclear

Figure 7 Calpain protease is activated in A3D8-induced cell death and involved in apoptosis-inducing factor (AIF) nuclear transloca- tion. (a) HEL cells were incubated with IgG1 (control) or A3D8 (3 mg/ml) in the absence or presence of calpain inhibitors ALLM (10 mM) and calpeptin (20 mM)orCa2 þ chelator BAPTA/AM (1 mg/ ml). Percentage (%) of cell death was estimated by trypan blue dye exclusion after 4 days of incubation. Values are means7s.d. of triplicate cultures. (b) HEL cells were transfected with negative control-small interfering RNA (siRNA) (NC-siRNA), CAPNS1- siRNA1 or CAPNS1-siRNA2. Six hours after electroporation, cells were cultured with IgG1 or A3D8 and cell death was estimated by trypan blue exclusion test at 3 days of treatment and compared with not transfected (control) cells. Values are means7s.d. of triplicate cultures. *Po0.05 compared with A3D8 alone or with A3D8 þ NC-siRNA. The inserted Western blot picture shows that calpain small subunit protein was diminished 3 days after electroporation in A3D8-treated cells transfected with CAPNS1- siRNA1 or with CAPNS1-siRNA2 as compared with NC-siRNA. (c) HEL cells were cultured for indicated days with 3 mg/ml of A3D8 or IgG1 (control). Total cell proteins were analysed by Western blotting with anti-calpain small subunit antibody. (d) Top: Nuclear fraction of HEL cells cultured with IgG1 (control) or A3D8 for 4 days in the absence or presence of calpain inhibitor ALLM (10 mM) was analysed by Western blotting with anti-AIF antibody. Membrane was striped and probed with anti-poly (ADP- ribose) polymerase-1 antibody as loading control for nuclear material. Bottom: Densitometry quantification of changes in nuclear AIF was derived from Western blots of three independent experiments with arbitrary unit corresponding to A3D8-treated cells. (e) Cytosolic fractions of cells transfected with NC-siRNA or CAPNS1-siRNA2 and treated for 4 days with A3D8 were analysed by Western blot with anti-AIF and anti-actin antibody.

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5748 condensation), membrane modifications (PS externali- Dou, 2000; Altznauer et al., 2004). However, a role for zation), specific mitochondrial (decrease in DCm), and calpain/Bax-p18/cytochrome c/caspase-dependent path- nuclear (chromatin condensation and DNA fragmenta- way is unlikely in A3D8-treated HEL cells in the tion) alterations. However, in view of the morphology of absence of cytochrome c release and caspase activation. the dying cells observed by electronic microscopy, Our data showing that calpain is involved in AIF A3D8-induced cell death can be classified as ‘apopto- translocation rather suggest that CD44-induced HEL sis-like PCD’ (Jaattela and Tschopp, 2003). Moreover, cell death is mediated by a calpain/AIF caspase- A3D8-induced HEL cell death appeared to be indepen- independent pathway. dent of caspases, a cysteine–protease family mainly Our results show a significant decrease in the involved in PCD, but rather caused by calpain and AIF, mitochondrial DCm associated with the release of AIF indicating that it was not a classical apoptosis. We from mitochondria and its translocation to nucleus. recently reported that A3D8 mAb induces apoptosis in Moreover, AIF activity characterized by a large-scale NB4 leukemic cells through caspase- and serine pro- (B50 kb) DNA fragmentation was detected in A3D8- tease-dependent pathways (Maquarre et al., 2005). It is treated HEL cells. Finally downregulation of AIF known that molecular partners of CD44 are various and expression by AIF-siRNA diminished A3D8-induced are likely to differ in different cell types explaining cell death suggesting that AIF is required for CD44- different signaling pathways and consecutively different induced in HEL cells. Recent studies have suggested that death pathways for a same agonist mAb. Our findings AIF is a central mediator of relevant experimental indicate that the cellular death response to CD44 models of cell death (Corbiere et al., 2004; Kang et al., ligation with A3D8 mAb depends on cellular context 2004; Liu et al., 2004a; Cheung et al., 2005; Ishitsuka and may involve caspase-dependent (in NB4 cells) as et al., 2005; Park et al., 2005). However, AIF is a janus well as caspase-independent pathways (in HEL cells). protein with a vital role in cells via its redox activity, and Interestingly, caspase-independent cell death has also a lethal role by participating to chromatinolysis and been described for two other mAb-activated cell surface chromatin condensation once translocated to the molecules, CD47 and CD99, in leukemic cells. However, nucleus (Lipton and Bossy-Wetzel, 2002; Vahsen et al., CD47- and CD99-induced cell death differs from CD44- 2004). This double activity of AIF may explain why induced HEL cell death as it occurs very rapidly after AIF-siRNA failed to completely block cell death in antigen ligation, involves an early reorganization of A3D8-treated HEL cells. Another possibility is that AIF cytoskeleton and is AIF-independent (Mateo et al., is not the alone effector of A3D8-induced cell death. 2002; Roue et al., 2003; Cerisano et al., 2004). Indeed, the phenotype of cell death described here was We were unable to detect the active forms of caspases- not identical to the previously identified AIF-induced 3, -7, -8 and -9 in A3D8-treated HEL cells. A plausible cell death involving delayed cytochrome c release and explanation for the absence of caspase-9 and -3 subsequent caspase cascade (Susin et al., 1999; Yu et al., activation in A3D8-treated HEL cells is the lack of 2002). Thus, it is possible that death effectors other than mitochondrial cytochrome c release. Although caspase- AIF contribute to the caspase-independent cell death specific inhibitors did not block A3D8-induced cell induced by A3D8. This point is an important issue that death the pan–caspase inhibitor Z-VAD-fmk, used at would be evaluated in a more specific study. high concentrations (X50 mM), was able to partially Recent observations indicated that AIF may be a key decrease cell death in A3D8-treated HEL cells. How- molecule in PARP-1-mediated caspase-independent cell ever, such high dosages of ZVAD-fmk are no longer death (Yu et al., 2002; Hong et al., 2004). Although the specific and may affect cysteine proteases other than mechanism of PARP-1-mediated release of AIF remains caspases, such as calpains or cathepsins (Schotte et al., elusive, it appears to be involved in apoptotic forms of 1999). Indeed, we detected an early activation of cell death. This type of PARP-1/AIF-mediated cell calpain, a calcium-dependent cysteine protease, in death is described to be particularly effective in cells very A3D8-treated HEL cells that was likely involved in cell resistant to cell death like cancer cells (Kang et al., death as two calpain inhibitors or transfection with 2004), neuronal cells (Wang et al., 2004) or cardiac calpain siRNA could partially suppress A3D8-induced myocytes (Chen et al., 2004). The major implication of cell death. Numerous studies have shown the potential AIF in A3D8-induced cell death has led us to examine role for calpains in necrosis and in PCD (Lu et al., 2002; the role of PARP-1 in this death. We show that 1,5- Liu et al., 2004b). Calpains have been shown to be often isoquinolinediol PARP-1 inhibitor partially inhibited associated with caspase activation in different apoptotic the disruption of DCm, AIF translocation and sub- pathways (Neumar et al., 2003; Altznauer et al., 2004). sequent cell death induced by A3D8. Moreover, the However, caspase-independent cell death mediated by accompanying NAD þ decrease supports the notion that calpains has also been described in several models, PARP-1 is activated in A3D8-treated HEL cells. It has including neuronal cell death (Lankiewicz et al., 2000) been recently shown that reactive oxygen species (ROS)- and vitamine-D-induced apoptosis in breast cancer cells mediated DNA damage triggers activation of PARP-1 (Mathiasen et al., 2002). Although calpain has a large and subsequent cell death (Yu et al., 2003; Kang et al., number of substrates, its exact role in cell death remains 2004). However, the ROS scavenger N-acetylcysteine unclear. Bax cleavage by calpain into p18/Bax appears (NAC) failed to inhibit A3D8-induced cell death (data to be involved in drug-induced apoptosis of leukemic not shown), indicating that ROS is not involved in DNA cells or in spontaneous neutrophil apoptosis (Gao and damage and PARP-1 activation in our model. The

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5749 mechanism by which CD44 ligation leads to PARP-1 recently reported that the mitochondrial apoptosis- activation in HEL cells remains to be determined as well induced channel (MAC), the putative cytochrome c as the precise role of PARP-1 activation in A3D8- release channel, is formed by oligomeric Bax and is induced cell death. suppressed by Bcl-2 or Bcl-XL (Dejean et al., 2005). As We have showed that PARP-1 activity as well as Bcl-XL is over-expressed in erythroleukemia cells calpain activity could mediate the release of AIF from (Benito et al., 1996) it is possible that the MAC/Bax the mitochondria, as well as cell death in A3D8-treated channel may be deficient to release cytochrome c in HEL cells. To date, there is no evidence that suggests a A3D8-induced HEL cells and that AIF is released functional link between PARP-1 and calpain activation through a specific channel distinct from the MAC in cell death. However, the fact that simultaneous channel. In any case, our data provide a unique cellular inhibition of PARP-1 and calpain had not additive model to characterize the possible existence of a specific cytoprotective effect (data not shown) suggests that AIF channel and to study the role of Bcl-2 family PARP-1- and calpain-mediated death pathways are not members in both cytochrome c and AIF release. independent of each other. As PARP-1 activation and In conclusion, our results demonstrate that CD44 calpain activation are involved in AIF translocation, it ligation with the anti-CD44 A3D8 mAb induces cell is tempting to speculate that both are required in this death in HEL cells, through a caspase-independent process but at distinct steps. mechanism involving calpain and AIF. The features The mechanisms regulating the release of specific observed in this kind of cell death indicate that CD44- mitochondrial proteins during cell death are not induced PCD involves signalling pathways that are completely understood (Donovan and Cotter, 2004; different from those involved in the classical caspase- Henry-Mowatt et al., 2004; Saelens et al., 2004). Our dependent apoptosis. This original death pathway open data showed that A3D8 treatment induced loss of DCm the way to a new therapy in the treatment of AML. and AIF release from mitochondria without cytochrome c release. The release of cytochrome c and AIF from mitochondria depends on two steps (Ott et al., 2002). Materials and methods The first step is the detachment of both proteins from the inner mitochondria membrane (IMM) into the IMS, Materials The mAb directed to human CD44 (clone A3D8) was the second is the translocation into the cytoplasm by purchased from Sigma Aldrich (Saint-Quentin-Fallavier, MOMP. The different nature of protein anchorage in France). The primary antibodies used for immunoblots were IMM may explain the fact that AIF and cytochrome c as follows: mouse anti-PARP, anti-caspase-7 and anti-cyto- have different pathways for mitochondrial release. It is chrome c mAbs from BD PharMingen (Le Pont de Claix, known that cytochrome c is associated with cardiolipin France), mouse anti-caspase-3 and anti-caspase-8 mAbs from at the IMM (Brown and Wuthrich, 1977) whereas AIF Alexis Biochemicals (San Diego, CA, USA), rabbit polyclonal is anchored in IMM by its N-terminal segment. Indeed, anti-caspase-9 from Cayman Chemical (Ann Arbor, MI, AIF needs to be cleaved for becoming a soluble and USA), mouse anti-calpain small subunit mAb from Chemicon apoptogenic protein (Otera et al., 2005). Recent studies International (Temecula, CA, USA) and rabbit polyclonal in isolated mitochondria have shown that AIF cleavage anti-actin antibody from Sigma. Two anti-AIF antibodies were used: anti-AIF mAb (E-1) from Santa Cruz Biotechno- is caspase-independent and involves specific cysteine logy (Santa Cruz, CA, USA) and rabbit polyclonal anti-AIF proteases like calpain I or cathepsins B, L and S (Polster from Sigma. The caspase inhibitors Z-VAD-fmk, Z-DEVD- et al., 2005; Yuste et al., 2005). Here, we show that fmk, Z-VEID-fmk, Z-VDVAD-fmk and Z-YVAD-fmk were calpain inhibition not only decreases A3D8-induced cell all purchased from ICN/Enzyme Systems Products (Liver- death but also inhibits AIF translocation suggesting that more, CA, USA). Calpeptin, ALLM, BAPTA/AM and 1,5- AIF release from mitochondria may be a direct isoquinolinediol were obtained from Calbiochem, San Diego, consequence of calpain activity in HEL cells. To our USA. Etoposide was purchased from Sigma. knowledge, this is the first report showing that calpain mediates AIF release from mitochondria in leukemic Cell culture and CD44 ligation cells. Whether nuclear AIF in A3D8-treated HEL cells is The human acute erythroleukemia HEL cells were cultured in the cleaved form of AIF and whether calpain is involved 5% CO2 and 95% humidified air atmosphere at 371C in RPMI in this cleavage remains to be determined. 1640 medium (GIBCO-BRL, Cergy Pontoise, France) with Different models have been proposed for MOMP 10% heat-inactivated fetal calf serum, 1 mM glutamine, 100 mg/ (Donovan and Cotter, 2004; Bras et al., 2005): one ml penicillin and 100 mg/ml streptomycin. Cells were split every 2 days at 2 105 cells/ml to ensure a logarithmic growth. For involves opening of mitochondrial permeability transi- Â CD44 ligation, cells were seeded at 2 Â 105/ml in culture tion pore (PTP) resulting in mitochondrial swelling and medium, and treated with various concentrations (0.5 to rupture of the OMM and nonspecific release of proteins 20 mg/ml) of A3D8 mAb (IgG1). Negative controls were from the IMS. Another model exclusively depends on incubated with the same concentrations of murine IgG1 Bcl-2 family members through the formation of (Coulter-Immunotech, Marseille-Luminy, France). autonomous Bax/Bak mitochondrial channels. An interaction of Bax with PTP has been also proposed to Cell cycle analysis induce MOMP. The release of AIF without cytochrome Cells (3 Â 105) were washed twice in cold phosphate-buffered c in A3D8-treated HEL cells suggests specific pathways saline 1 Â (PBS) and fixed in ethanol 70% in PBS (30 min in for the release of both mitochondrial factors. It has been ice). Cells were harvested and incubated in 2 ml of solution

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5750 containing 1 mg/ml RNase, PBS and 1 ml/ml of propidium PolyScreen membranes (NEN, Paris, France) and processed iodirde (PI) for 30 min at 371C in the dark. Cells were analysed for western blot analysis as described previously (Hafid- by flow cytometry (FACS calibur, Becton Dickinson, Le Pont Medheb et al., 2003). Nonspecific binding sites were blocked de Claix, France). by pretreating membranes overnight at 41C with 5% unfatted milk in Tris-buffered saline (TBS)/0.05% Tween 20 Cell viability analysis and apoptosis assays TBS with 0.1% tween (TBS-T). Membranes were incubated Cells (2 Â 105 cells/ml) were treated with various concentra- 1 h at room temperature with primary antibodies in TBS-T tions of A3D8 and, at the appropriate time, viable cells were and then incubated with anti-mouse or anti-rabbit antibody counted every 24 h for 4 days using the trypan blue dye conjugated with horseradish peroxidase (HRP) for 1 h at exclusion assay. Apoptosis was determined by staining with a room temperature. Each step was followed by two washes for mixture of EB and AO as previously described (Poindessous- 15 min in TBS-T. The protein bands were detected as Jazat et al., 2002). In some experiments, apoptosis was described in the ECL protocol (NEN) using Reflection analysed by annexin V/p(PI) staining with a commercially autoradiography film. available kit (BD PharMingen) as recommended by the manufacturer. Briefly, 1 Â 105 cells were collected, washed in Immunocytochemistry and confocal microscopy cold PBS and incubated in binding buffer containing Cells were fixed with 3% paraformaldehyde in PBS for 30 min fluorescein isothiocyanate (FITC)-labeled annexin V and PI at 41C, and washed three times with PBS. After permeabiliza- for 15 min at room temperature in the dark. Samples were then tion with 0.05% triton X-100 for 5 min at room temperature, analysed by flow cytometry with a FACSCalibur (Becton- cells were incubated with rabbit polyclonal anti-AIF antibody Dickinson). To detect DNA strand breaks, the TUNEL assay (Santa Cruz; H-300). Cells were stained with FITC-conjugated was performed using the Apoptosis Detection System from Alexa goat anti rabbit IgC 488. After incubation in 0.5% PI Roche Diagnostics (Meylan, France). Apoptotic DNA frag- solution for 5 min, cells were examined by conventional or ments were end-labeled in situ with fluorescein-12-dUTP by confocal fluorescence microscopy (Leica Microsystemes, Rueil terminal deoxynucleotide transferase (TdT) and cells were then Malmaison, France) analysed on FACSCalibur. Pulse-field inversed gel electrophoresis Analysis of mitochondrial DCm Nuclear DNA from lysed cells (treated with protease K Variations of the DCm were determined by a previously and RNase) was prepared from agarose plugs (1 Â 106 cells) described procedure (Petit et al., 1995). Briefly, samples of digested twice with proteinase K (1 mg/ml; 501C; 12 h) in NDS 1 106 cells were washed twice in PBS, incubated with 30 nM Â buffer (0.5 M EDTA, 10 mg/ml lauroyl sarcosine), washed in 1 DiOC6 (3) for 10 min at 37 C and analysed by flow cytometry 0.5 Â TBE, followed by electrophoresis in a Bio-RadCHEF- using the FACSCalibur cytometer (FL-1 channel). DR II (Richmond, CA, USA) equipment (1% agarose; 0.5 Â TBE; 200 V; 24 h; pulse wave 60 s; 1201 angle). Preparation of nuclear extracts Cells (1 Â 106) were centrifuged, washed twice in cold PBS, Determination of NAD þ gently resuspended in a first lysis buffer (20 mM N-2-hydroxyl Cellular NAD þ levels were determined by the enzymatic piperazine-N0-2-elthane sulfonic acid (HEPES) pH 8, 10 mM cycling method using alcohol dehydrogenase as described in KCl, 0.2% NP40, 0.1 mM ethylene diamine tetraacetic detail by Zong et al. (2004). The concentration of NAD þ was acid(EDTA), 1 mM dithiothreitol (DTT), 10% glycerol, normalized by protein content as determined by the Bradford 100 mM phenylmethylsulfonyl fluoride (PMSF), 10 mg/ml leu- peptin, 2 mg/ml aprotinin) and then centrifuged for 5 min at assay. 350 g. Pellet was energically resuspended in a second lysis buffer (20 mM HEPES pH 8, 10 mM KCl, 350 mM NaCl, Cell transfection and small interfering RNAs 0.1 mM EDTA, 1 mM DTT, 20% glycerol, 100 mM PMSF, HEL cells were transfected with a siRNA doubled-stranded 10 mg/ml leupeptin, 2 mg/ml aprotinin). After centrifugation at oligonucleotide designed to interfere with the expression of 20 000 g for 15 min, supernatant (nuclear fraction) was human AIF (sense strand: 50-CCGGUCCCAGGCAACUUG- removed and stored at À801C. 30, Proligo, Boulder, CO, USA). An irrelevant siRNA oligonucleotide against CD9 (sense strand: 50-ACC Preparation of cytosol for measurement of cytochrome c and TCCTCCAGCTCGCTTA-30) was used as a control for AIF apoptosis-inducing factor (Ctrl-siRNA). For calpain silencing, we used two stealth For cytosol isolation, a modified technique based on digitonine siRNA 25-bp duplexes corresponding to calpain small subunit subcellular fractionation was used (Adrain et al., 2001). (CAPNS1) designed and purchased from Invitrogen (Paisley, Briefly, 1 Â 106 cells were centrifuged, washed twice in PBS UK). The sense strand of CAPNS1 primer 1 was AAACCA and incubated with a lysis buffer (250 mM sucrose, 70 mM KCl, UCAGUCUUCAGAUCAGGGU and of CAPNS1 primer 2 137 mM NaCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4 pH 7.2, was AGAGAUGCUCAUUCAGGUGGAACCC. The con- 100 mM PMSF, 10 mg/ml leupeptin, 2 mg/ml aprotinin, 200 mg/ trol stealth siRNA for CAPNS1 was a recommended negative ml digitonin), for 5 min on ice. After centrifugation at 1000 g control (NC-siRNA) purchased from Invitrogen. For transfec- for 5 min at 41C, the supernatant (cytosolic fraction) was tion, cells (10 Â 106 cells in 400 ml of RPMI) were electro- removed and stored at À201C. porated at room temperature with 0.5 mM of siRNA using the Pulser apparatus (Bio-Rad, Ivry, France) with con- Western blot analysis ditions of 300 V and 500 microfarads. Cells (5 Â 106) were washed in PBS, lysed with 300 ml buffer containing 20 mM Tris-HCl pH 7.0, 0.5% SDS and 1% Statistical analysis benzonase (Sigma Aldrich) and then inactivated at 1001C for Data are expressed as means7s.d. of three independent 3 min. Protein lysates were separated by 8 or 12% SDS– experiments and statistically analysed using Student’s t-test. polyacrylamide gel electrophoresis (SDS-PAGE), blotted onto Values of Po0.05 were considered as significant.

Oncogene CD44-induced cell death in erythroleukemia cells CArtuset al 5751 Acknowledgements sur le Cancer (contract No 4812), Fondation de France and Pasteur-Weizmann Scientific Council (to SAS). RSM was This work was supported by grants from Association NRB- supported by a PhD fellowship from Fondation Hariri. CD Vaincre le Cancer (to JR-L), Association pour la Recherche holds a postdoctoral fellowship from Fondation de France.

References

Adrain C, Creagh EM, Martin SJ. (2001). Embo J 20: Liu T, Brouha B, Grossman D. (2004a). Oncogene 23: 39–48. 6627–6636. Liu X, Van Vleet T, Schnellmann RG. (2004b). Annu Rev Altznauer F, Conus S, Cavalli A, Folkers G, Simon HU. Pharmacol Toxicol 44: 349–370. (2004). J Biol Chem 279: 5947–5957. Lockshin RA, Zakeri Z. (2004). Oncogene 23: 2766–2773. Benito A, Silva M, Grillot D, Nunez G, Fernandez-Luna JL. Lu T, Xu Y, Mericle MT, Mellgren RL. (2002). Biochim (1996). Blood 87: 3837–3843. Biophys Acta 1590: 16–26. Bras M, Queenan B, Susin SA. (2005). Biochemistry (Moscow) Maquarre E, Artus C, Gadhoum Z, Jasmin C, Smadja-Joffe F, 70: 231–239. Robert-Lezenes J. (2005). Leukemia 19: 2296–2303. Broker LE, Kruyt FA, Giaccone G. (2005). Clin Cancer Res Mateo V, Brown EJ, Biron G, Rubio M, Fischer A, Deist FL 11: 3155–3162. et al. (2002). Blood 100: 2882–2890. Brown LR, Wuthrich K. (1977). Biochim Biophys Acta 468: Mathiasen IS, Sergeev IN, Bastholm L, Elling F, 389–410. Norman AW, Jaattela M. (2002). J Biol Chem 277: Cande C, Vahsen N, Garrido C, Kroemer G. (2004). Cell 30738–30745. Death Differ 11: 591–595. Miramar MD, Costantini P, Ravagnan L, Saraiva LM, Cerisano V, Aalto Y, Perdichizzi S, Bernard G, Manara MC, Haouzi D, Brothers G et al. (2001). J Biol Chem 276: Benini S et al. (2004). Oncogene 23: 5664–5674. 16391–16398. Charrad RS, Gadhoum Z, Qi J, Glachant A, Allouche M, Neumar RW, Xu YA, Gada H, Guttmann RP, Siman R. Jasmin C et al. (2002). Blood 99: 290–299. (2003). J Biol Chem 278: 14162–14167. Charrad RS, Li Y, Delpech B, Balitrand N, Clay D, Jasmin C Nicholson DW. (1999). Cell Death Differ 6: 1028–1042. et al. (1999). Nat Med 5: 669–676. Otera H, Ohsakaya S, Nagaura Z, Ishihara N, Mihara K. Chen M, Zsengeller Z, Xiao CY, Szabo C. (2004). Cardiovasc (2005). EMBO J 24: 1375–1386. Res 63: 682–688. Ott M, Robertson JD, Gogvadze V, Zhivotovsky B, Orrenius Cheung EC, Melanson-Drapeau L, Cregan SP, Vanderluit JL, S. (2002). Proc Natl Acad Sci USA 99: 1259–1263. Ferguson KL, McIntosh WC et al. (2005). J Neurosci 25: Park YC, Jeong JH, Park KJ, Choi HJ, Park YM, Jeong BK 1324–1334. et al. (2005). Life Sci 77: 2059–2070. Corbiere C, Liagre B, Terro F, Beneytout JL. (2004). Cell Res Petit PX, Lecoeur H, Zorn E, Dauguet C, Mignotte B, 14: 188–196. Gougeon ML. (1995). J Cell Biol 130: 157–167. Cregan SP, Dawson VL, Slack RS. (2004). Oncogene 23: 2785– Poindessous-Jazat V, Augery-Bourget Y, Robert-Lezenes J. 2796. (2002). Leukemia 16: 233–243. Dejean LM, Martinez-Caballero S, Guo L, Hughes C, Teijido Polster BM, Basanez G, Etxebarria A, Hardwick JM, Nicholls O, Ducret T et al. (2005). Mol Biol Cell 16: 2424–2432. DG. (2005). J Biol Chem 280: 6447–6454. Donovan M, Cotter TG. (2004). Biochim Biophys Acta 1644: Ponta H, Sherman L, Herrlich PA. (2003). Nat Rev Mol Cell 133–147. Biol 4: 33–45. Gadhoum Z, Delaunay J, Maquarre E, Durand L, Lancereaux Roue G, Bitton N, Yuste VJ, Montange T, Rubio M, V, Qi J et al. (2004a). Leuk Lymphoma 45: 1501–1510. Dessauge F et al. (2003). Biochimie 85: 741–746. Gadhoum Z, Leibovitch MP, Qi J, Dumenil D, Durand L, Saelens X, Festjens N, Vande Walle L, van Gurp M, van Loo Leibovitch S et al. (2004b). Blood 103: 1059–1068. G, Vandenabeele P. (2004). Oncogene 23: 2861–2874. Gao G, Dou QP. (2000). J Cell Biochem 80: 53–72. Schotte P, Declercq W, Van Huffel S, Vandenabeele P, Beyaert Ghaffari S, Smadja-Joffe F, Oostendorp R, Levesque JP, R. (1999). FEBS Lett 442: 117–121. Dougherty G, Eaves A et al. (1999). Exp Hematol 27: 978– Song G, Liao X, Zhou L, Wu L, Feng Y, Han ZC. (2004). 993. Leuk Res 28: 1089–1096. Hafid-Medheb K, Augery-Bourget Y, Minatchy MN, Hanania Sun XM, MacFarlane M, Zhuang J, Wolf BB, Green DR, N, Robert-Lezenes J. (2003). Blood 101: 2575–2583. Cohen GM. (1999). J Biol Chem 274: 5053–5060. Henry-Mowatt J, Dive C, Martinou JC, James D. (2004). Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snow BE, Oncogene 23: 2850–2860. Brothers GM et al. (1999). Nature 397: 441–446. Hong SJ, Dawson TM, Dawson VL. (2004). Trends Pharmacol Vahsen N, Cande C, Briere JJ, Benit P, Joza N, Larochette N Sci 25: 259–264. et al. (2004). Embo J 23: 4679–4689. Ishitsuka K, Hideshima T, Hamasaki M, Raje N, Kumar S, Wang H, Yu SW, Koh DW, Lew J, Coombs C, Bowers W Podar K et al. (2005). Oncogene 24: 5888–5896. et al. (2004). J Neurosci 24: 10963–10973. Jaattela M, Tschopp J. (2003). Nat Immunol 4: 416–423. Yu SW, Wang H, Dawson TM, Dawson VL. (2003). Neurobiol Johnson DE. (2000). Leukemia 14: 1695–1703. Dis 14: 303–317. Joza N, Susin SA, Daugas E, Stanford WL, Cho SK, Li CY Yu SW, Wang H, Poitras MF, Coombs C, Bowers WJ, et al. (2001). Nature 410: 549–554. Federoff HJ et al. (2002). Science 297: 259–263. Kang YH, Yi MJ, Kim MJ, Park MT, Bae S, Kang CM et al. Yuste VJ, Moubarak RS, Delettre C, Bras M, Sancho P, (2004). Cancer Res 64: 8960–8967. Robert N et al. (2005). Cell Death Differ 12: 1445–1448. Lankiewicz S, Marc Luetjens C, Truc Bui N, Krohn AJ, Poppe Zada AA, Singh SM, Reddy VA, Elsasser A, Meisel A, M, Cole GM et al. (2000). J Biol Chem 275: 17064–17071. Haferlach T et al. (2003). Oncogene 22: 2296–2308. Leist M, Jaattela M. (2001). Nat Rev Mol Cell Biol 2: 589–598. Zong WX, Ditsworth D, Bauer DE, Wang ZQ, Thompson Lipton SA, Bossy-Wetzel E. (2002). Cell 111: 147–150. CB. (2004). Dev 18: 1272–1282.

Oncogene